mjh Posted December 25, 2012 Share #181 Posted December 25, 2012 Advertisement (gone after registration) If you've ever been 'interviewed' for LFI, your words will be translated into German . . . and then back into English again for the English edition. A disturbing experience! I will do my best to make sure that won’t happen again. Link to post Share on other sites More sharing options...
Advertisement Posted December 25, 2012 Posted December 25, 2012 Hi mjh, Take a look here M-240 "Seeking Light". I'm sure you'll find what you were looking for!
tgm Posted December 25, 2012 Share #182 Posted December 25, 2012 On the screen of the laptop computer of a Leica representative at a 2-day workshop in Solms. He was very proud that their new camera (with premature firmware!) beats the current high-end Canon EOS model with regard to high-ISO noise. But then, the EOS 1DX can go far beyond ISO 6400/39° while the Leica M cannot. And those sample shots were taken in fairly good light—i. e. moderate contrast and no seriously underexposed shadows. So the fact that the M looked better than the 1DX in these particular high-ISO shots does not necessarily mean it will look better under all possible circumstances. More testing required. But what we saw definitely is a good start ... This sounds indeed very promising. For me 6400 ASA is already very attractive. Actually, it is not clear to me what this very high ISO numbers of Canon 1DX up to 100.000 ASA or Nikon D4 even 200.000 ASA mean. The sensor has a native sensitivity, usually around 200 ASA for common full frame sensors. Higher ISO means underexposure and a compensation by the camera software. So if 6400 ASA of the Leica M is not sufficient, it should be possible to underexpose and try to correct in Photoshop. In principle I would not expect that such results should differ a lot from shots taken with extremely hight ISO numbers, but it is much more post processing needed. Or do I make a mistake? Thomas Link to post Share on other sites More sharing options...
pico Posted December 25, 2012 Share #183 Posted December 25, 2012 Higher ISO means underexposure and a compensation by the camera software. So if 6400 ASA of the Leica M is not sufficient, it should be possible to underexpose and try to correct in Photoshop. That is an interesting problem. I am not sure one can make up for such underexposure in Photoshop. Does not the camera amplify the signals to achieve higher ISO? Isn't that different? Link to post Share on other sites More sharing options...
jonoslack Posted December 25, 2012 Share #184 Posted December 25, 2012 If you've ever been 'interviewed' for LFI, your words will be translated into German . . . and then back into English again for the English edition. A disturbing experience! I will do my best to make sure that won’t happen again. Hi Michael personally - no worries - and I was certainly not 'mis-quoted'. But it did all sound a lot more considered and serious than I felt! But I agree, it's a good principle! all the best Link to post Share on other sites More sharing options...
wildlightphoto Posted December 25, 2012 Share #185 Posted December 25, 2012 ... If you've ever been 'interviewed' for LFI, your words will be translated into German . . . and then back into English again for the English edition. A disturbing experience! That's been my experience too. The phrase 'lost in translation' took on new meaning for me Link to post Share on other sites More sharing options...
luigi bertolotti Posted December 26, 2012 Share #186 Posted December 26, 2012 That is an interesting problem. I am not sure one can make up for such underexposure in Photoshop. Does not the camera amplify the signals to achieve higher ISO? Isn't that different? I seem to have read somewhere that the ultra high ISO of some dslrs are achieved not only by signal amplification, but also by "coupling" squares of pixels (2x2, 3x3, 4x4 ?), lowering the effective resolution and,of course, applying different in camera processing to the bitmap Link to post Share on other sites More sharing options...
Erik Gunst Lund Posted December 26, 2012 Share #187 Posted December 26, 2012 Advertisement (gone after registration) 'Coupling' pixels is called Binning. The Nikon D1 had 2x2 binning and the D1X had 1x2 binning. Link to post Share on other sites More sharing options...
mjh Posted December 26, 2012 Share #188 Posted December 26, 2012 That is an interesting problem. I am not sure one can make up for such underexposure in Photoshop. Does not the camera amplify the signals to achieve higher ISO? Isn't that different? Yes and no. The camera does amplify the sensor signals to compensate for the underexposure at higher ISO settings, but only up to a point. For most cameras there is a point where it switches from increasing the analogue gain to digital amplification, i.e. multiplication. With modern CMOS sensors this switch happens around ISO 1600. At ISO 6400 and above were are within an ISO range where any increase in ISO is achieved by multiplication. The reason for this switch is that depending on the contribution of the various sources of noise, analogue amplification is only beneficial for moderately high ISO levels; there is a point above which amplification ceases to offer any advantage over multiplication. Link to post Share on other sites More sharing options...
Rick Posted December 26, 2012 Share #189 Posted December 26, 2012 Michael, can you explain CMOSIS' patented two part signal processing system that takes place at the level of the A/D converters at the bottom of each column on the sensor? They call it Digital Correlated Double Sampling. Guy Meynants of CMOSIS claims that it reduces noise very effectively. How would this patented design be different (better) than how other manufacturers design their sensors? Will this give the CMOSIS chip a noise advantages over other chips? He also claims it is used in the processing of 1080p video. It sounds like CMOSIS used very good and fast A/D converters and is able to double sample information by using fast switching that these A/D converters would allow. This double switching and the fast A/D converters also improves video but, I'm not sure how. Link to post Share on other sites More sharing options...
mjh Posted December 26, 2012 Share #190 Posted December 26, 2012 Double Correlated Sampling is a fairly standard technology used for reducing the fixed-pattern noise inherent in CMOS sensors. You sample the output of a sensor pixel once in a known state (i.e. when its capacitator is supposedly empty) and then again for the actual read-out at the end of the exposure. By subtracting the first reading from the second, any non-uniformity is eliminated. Digital Double Correlated Sampling would digitise each sample and subtract the digised values, or at least that’s my understanding. Link to post Share on other sites More sharing options...
Rick Posted December 26, 2012 Share #191 Posted December 26, 2012 Michael, Thanks for the comments. On page 66 of the latest LFI Meynants makes it sound like they are doing something different than other manufacturers. By they way, I enjoyed your latest two articles in the last LFI about, "Colour Dynamics" and the article, "Mastering the M Monochrom." That one is going to take me a couple of reads through. Rick Link to post Share on other sites More sharing options...
leica5 Posted December 26, 2012 Share #192 Posted December 26, 2012 Posted pictures remind me of zeiss lens foot print(good),,my newly purchased Sony RX1 is definitely a keeper. Sami Link to post Share on other sites More sharing options...
algrove Posted December 26, 2012 Share #193 Posted December 26, 2012 Posted pictures remind me of zeiss lens foot print(good),,my newly purchased Sony RX1 is definitely a keeper. Sami I have read about it, the RX1. How do you get on with having, but one fixed lens? If you own an M9, how do the images compare? Thanks. Link to post Share on other sites More sharing options...
Recommended Posts
Archived
This topic is now archived and is closed to further replies.